If you run a subscription business, your cancellation flow is either making you money or quietly costing you your best customers. Most operators never find out which, because they never test it. They install Recharge or Skio, turn on the default save offers, and move on.

We didn't move on. We tested it, one offer at a time, with a rapidly scaling supplement client, and found the single lever that actually stops churn before it happens.

When a subscriber clicked cancel, we didn't show them a menu of five ways to stay. We tested offers in isolation, first the offer and nothing else, to find out which one actually moved someone off the cancel button on its own. A free gift. A pause. A skip. A delay on the next shipment. A discount on upcoming orders.

The free gift barely moved the needle. Pause and skip did better than nothing, but not by much. Delay was closer to a formality than a save.

The discount won by a mile. Not a discount forever. A discount on the next two to three orders only.

Here's the part most operators would get wrong if they tried to copy this. They'd assume the discount worked because it made the product cheaper, so they'd apply it indefinitely to be safe. That's the fastest way to turn a great result into a bad one.

The discount doesn't work because it's cheap. It works because it buys time. A subscriber who hits cancel in month two or three hasn't built a routine yet. They're still deciding whether this is worth remembering to reorder, worth the money, worth the friction. A permanent price cut doesn't solve that. It just makes it cheaper to stay undecided. A capped discount does something different. It removes the price objection just long enough for the product to become a habit instead of a decision. By the third discounted order, the customer isn't evaluating the product anymore. They're just using it. And once someone is using your product instead of thinking about it, price stops being the reason they'd leave.

That's why the cap matters as much as the discount. Two to three orders, then it ends. Open-ended discounting trains your subscribers to expect a lower price forever and erodes your margin with nothing to show for it, except a customer who now churns at the discounted price instead of the full one. The two-to-three order window is deliberate. It's built to expire right around the point where habit takes over from hesitation.

If your cancellation flow is built around variety, more offers, more paths, more ways to say wait, don't go, you're solving the wrong problem. The question isn't how many offramps you can give someone. It's which single offer, shown first, buys you enough time for the routine to form.

For us, on this client, it was a short, capped discount. Your number might be different. But if churn is the problem, this is where to start testing.

See you next week.